Personalization tailors a customer's or audience's experience based on insights about their behaviors and needs. Its goal is to improve the customer experience (CX) and drive business outcomes.
Companies that grow faster drive 40 percent more of their revenue from personalization than their slower-growing counterparts.
—McKinsey & Company, November 2021
Personalization needs to go beyond improving digital experiences; it needs to avoid delivering a negative experience that could frustrate or alienate someone. We’ve all received marketing or service messages that clearly communicated the wrong thing or provided the wrong content.
A well-personalized experience is great, but a poorly personalized one can damage customer relationships. At best, personalization should help create authentic, relevant, and context-aware digital experiences. Perhaps more importantly, it helps to prevent damage to your brand from a poorly timed, discordant, or offensive message.
One way to tackle personalization is to break it down into three basic elements.
Context helps us understand what customers experience and whether it meets their expectations. Context also informs what content or product recommendations are delivered. Context and the intelligence it provides must come quickly (ideally within milliseconds) for a person to view their experience and recommendation as relevant. The process to maximize contextual relevance can be broken into three steps:
Customer data comes from a variety of sources—website tags, apps, transactional systems, loyalty programs, and customer service software—and can be made available for a common, trusted customer profile. If there is even a millisecond's delay in collecting this data, it will become outdated and not be useful. In addition, if it violates consent and privacy considerations, the data must not be used. It probably should have not been collected in the first place.
Data that has been gathered and processed must be analyzed to form a real-time view of the customer’s context. With this view of who the customer is and what the customer expectation is, the customer experience can be personalized and orchestrated using a mixture of rules and algorithms to make it the best possible experience. By aggregating data across all customer touchpoints, you can use machine learning (ML) to understand common journeys, determine intent, and maximize desired outcomes.
Delivering the actual experience should not be underestimated. Delivery into the right channel(s), at the right time and with the right content, at scale, requires automation. Additionally, if there is not enough information to personalize an experience (i.e. you don’t have enough knowledge about who a person is) it’s essential to gracefully fall back to a default experience.
Context—including device, language, channel, journey stage, time, and location—and intelligence is needed to personalize content. Building, managing, and delivering a ready supply of engaging content also requires the right strategies, processes, and technologies.
You can improve existing content simply by improving the metadata and providing centralized access so its easier to find. You can also use analytics to improve how your content is used and its impact. You can also automate new content creation with an accelerated and decentralized approach that's able to manage for quality and consistency.
The two most important questions of any personalization strategy are “What are we trying to achieve?” and “How will we know when we’ve achieved it?” If there are no answers, the process should stop until all stakeholders can agree on the desired result.
The desired outcomes of any personalization strategy can vary widely but usually fall into three buckets:
Many personalization strategies start out focused on growth and are led by marketing teams. However, there is a maturity curve for personalization strategies, where a fourth outcome—customer loyalty and customer lifetime value—defines the approach.
Unfortunately, optimizing a personalization offering for outcomes vastly outstrips someone's decision-making powers as well as most rule-based systems. To optimize a personalization strategy for multiple outcomes requires AI and ML. AI/ML algorithms are needed to handle the unfathomably large amount of data and calculations. They can do so in a fraction of the time that it takes a human.
As optimization complexity grows, the sophistication of the algorithms must increase must be accompanied by mature data science skills and governance to keep the program in line with strategic and regulatory requirements.
While data science tools are the most robust for personalization efforts, they’re out of reach for most marketing professionals. But those same professionals need the right strategy to provide personalized experiences and right technologies that enable those experiences. Marketing professionals need personalization tools that support their daily tasks with the level of automation needed to personalize at scale. Examples of these jobs include recommendations, orchestration, targeting, testing, and measuring. They are described in more detail below:
Driving recommendations and suggestions is a large part of why marketing teams look to personalize their efforts. It’s what powers the extremely popular “You may also like…” sections of websites. These solutions manage a large inventory of content and leverage rules and algorithms to push the right content into various channels and platforms (web, mobile, email, search results, etc.). They also ensure that the best recommendations are consistent across channels.
Forgot that product in your cart? Abandoned that service request form? Personalizing multiple campaigns across different channels requires orchestration. Orchestration is much more than management. It manages the connections and dependencies between the systems that are used to communicate with a customer, not just for marketing messages, but also other areas of the business, such as customer service.
Audience and customer targeting is part of any successful personalization strategy. An experience that works for one group of people isn’t going to work for everyone. Which is why targeting specific audiences will generate higher returns than taking a one-size-fits-all approach. Marketing teams can leverage the dynamic templates and filters found within personalization tools and make adjustments based on the right attributes to deliver the right experiences at a 1-1 level.
Both A/B and multivariate testing are needed to determine the most impactful personalization strategy. Tests can be done on almost any type of personalized experience. This includes how a website is tailored and what email content will generate the highest engagement levels with certain customers. No matter what is being tested, the approach should be rigorous, going beyond hunches to data-backed decisions.
In addition, testing should do the following:
As the Peter Drucker saying goes, “You can only manage what you can measure.” Measurement and reporting solutions can slice and dice data at multiple levels—within and across channels, across campaign variants, outcome attribution, opportunity discovery, and much more. Once the results are available, marketers should share them with stakeholders so that everyone is on the same page, ensuring that the right steps are taken and any needed course corrections are made.
For personalization to work, the largest number of people need to have access to the right tools and technologies. The days of IT implementing these types of strategies is no longer feasible because of the speed required. The good news is that there are many user-friendly and intuitive personalization tools available.
Personalization efforts can be partially supported across various marketing and CRM systems. However, a true personalization tool is an application that can be used specifically for real-time personalization. It collects and analyzes digital behaviors to deliver personalized, cross-channel customer experiences in real time.
Explore Oracle Infinity Behavioral Intelligence’s personalization capabilities